Hal 9000 Analyzes 28,000 Year Old Rock Art
Archaeologists in Australia use machine learning to study ancient petroglyphs.
A new study tests the ability of AI machine learning to help analyze cave art in Australia. In particular, archaeologists are looking for new insights into how artistic styles changed with time.
By showing a computer program 1,000 images, researchers were able to teach it to recognize different images and, importantly, tell how similar they were to each other. This provides us with a mathematical model that shows the precise relationship between two images in a measurable way that removes at least some human biases.
Essentially, the model found that two types of figures we know are closely related in time are also closely related in style. The important thing here though is that this similarity is no longer, “These look similar to me” it is “The graph shows how similar they are.”
This does not mean that no bias is present in the results. Racial bias in facial recognition software has shown us that our own biases still find a way in. However, this is still a huge step forward for archaeological research.
Continued on page 23
Archaeology Helps Save Migrant Lives
As migrants are pushed to take more dangerous routes to cross the southern boarder of the U.S., risk of death due to exposure becomes a serious concern.
A geographer is using archaeological and military models for estimating energy cost in an attempt to help save lives. Using these tools, humanitarian groups can calculate where the greatest need for water is
Continued on page 11
Q: How much are you paid to hide the existence of aliens?
A: Wait, someone is getting paid?!
No comments:
Post a Comment